Are accident and emergency attendances increasing?

Pressures on accident and emergency (A&E) departments hit the headlines last week, with the Prime Minister and Leader of the Opposition trading statistics across the despatch box at Prime Minister’s Question Time. But what are the facts about A&E attendances?

Figure 1 shows trends in attendances in English NHS A&E units over the past 26 years. The topline trend shows that for 15 years – from 1987/8 to 2002/3 – attendances were essentially unchanged at around 14 million per year. But in 2003/4 they jumped – by nearly 18 per cent – to 16.5 million, and rose to 21.7 million by 2012/13. This is an overall rise of around 7.5 million (a 50 per cent increase) over the past decade.

Figure 1: Annual attendances in English A&E units: 1987/8 to 2012/13

Data source: Department of Health, Total time spent in accident and emergency (pre-2011/12 Q2) and NHS England, A&E waiting times and activity (current)

While the Secretary of State for Health, Jeremy Hunt, has suggested that changes to the GP contract around 2004 are to blame for the rise in A&E activity, the facts about accident and emergency workload statistics are not straightforward.

As Figure 1 also shows, in 2003/4 – when the large increase in attendances started – there was a change in the data series. Until 2003/4, statistics on A&E attendances included ‘major’ A&E units only. But around this time more, smaller units – including walk-in centres (WiCs) and minor injuries units (MIUs) – were introduced with the intention of diverting less serious emergency cases away from the larger, more expensive A&E departments, and the statistical collection was changed to record attendances separately for ‘type 1, 2 and 3’ units. Type 1 essentially reflecting major A&E units and types 2 and 3 defined as the smaller, walk-in and minor injuries units, together with specialist emergency departments.

So, much of the increase in 2003/4 was due to previously unrecorded attendances now being collected, but also additional – but less serious – work being carried out in the new units. From 2003/4 to 2012/13, attendances in type 1 units have remained more or less unchanged. It is attendances in type 2 and 3 units that account for all the increase.

While the NHS has experienced a phenomenal increase in accident and emergency workload over the past decade, over the past 30 months this increase has started to level off. Figure 2, for example, shows weekly attendances between November 2010 and April 2013 together with a trend line. The trend increase over the whole period is around 3.5 per cent – about 1.3 per cent per year.

Figure 2: Weekly attendances in A&E units: November 2010 to April 2013

Data source: NHS England, Weekly SitReps, 2013-14

Over the past decade there has certainly been a huge increase in work. But this is partly attributable to the changes in statistical collection and a degree of ‘supply induced demand’ as new routes into emergency care (WiCs, MIUs) opened. In addition, the possible substitution of some types of care (a visit to the GP) by others (a visit to a minor injuries unit) may have increased A&E attendance statistics, but given that WiCs and MIUs are generally not open out-of-hours, it is not easy to see how the rising workload of these services is linked to changes in GP out-of-hours arrangements, as implied by the Secretary of State for Health.

The increase in A&E workload has slowed in recent years, so the question now is not so much how to slow the speed of the increase, but how close to capacity the system is. The recent figures showing that the target that 95 per cent of patients should not wait longer than four hours in A&E is being breached on an increasingly regular basis is indicative of capacity problems amongst other things. Slowly raising the temperature does eventually boil the frog.

Comments

Looks to me like planned follow-up attendances were included at least pre-2003-04 but excluded at least from 2011-12 on. certainly the 16.5 million in 2003-04 includes planned follow-up. so the changes in data definition are causing even more problems. if I am right, then A&E attendances have probably gone up more in the later years than you indicate, but I do not have a figure for total follow-ups, nor breakdown between planned and unplanned follow-ups.
ps i think the word verification is a little inappropriate.

A very interesting article. Aren't there two issues it doesn't address though?

1. Haven't GP attendances gone up substantially over this period? So WICs and the rest not only have not reduced type 1 A&E activity, they haven’t substituted for primary care consultations

2. Don’t some A&E departments have within them type 2 units? Not all type 2 units are off site, so total A&E activity has continued to go up and type 1 activity has not gone up as planned, indeed from the graph it appears to continue to rise.

I think you used the word admission rather than attendance in para 5. An important difference which is sometimes lost in debate about GP access, so-called inappropriate attendances and the much bigger issue of changes in acute need.

In terms of waiting times and heat in waiting rooms/ A&E itself, attendances are important, but in terms of overall heat in acute system, pressure on beds and overall costs, admissions and bed days are surely the important issue rather than attendances.

I know KF has done lots on this too.

Thank you for clarifying. To Mark's point, GP consults have gone up too so no substitution for WIC activity there, but have the WICs taken some of the demand which would have gone to A&E and therefore helped? Your data suggests not. Why would the previous decade trend not have continued with very slow growth? If so, we have shown supply induced demand once again, or as the saying goes, "Build it, and the people will come."

If anyone doubts that market models do not apply well to the NHS, the notion that “supply led demand” is a bad thing should convince them. The commercial world employs a sales force precisely for the purpose of increasing demand to consume the available supply.

In the world of the NHS we have the problem of inequality of access to healthcare; for example people from a deprived locality are less likely to receive a timely diagnosis of cancer or to have a joint replaced. The emergency department, of whatever type, represents a facility where they can compete on an even playing field so an increase in supply for them is a good thing. So perhaps John Appleby is showing us an example of temporarily easing the problem of supply-limited demand.

Thanks for these comments. Here are some responses:

Sean: yes, possible problem re 'planned cases'...but we are reasonably assured by DH stats people that none of the 'subsequent' attendance data in either graph includes planned subsequent attendances. We have trawled through guidance docs as far as possible and these seem to confirm this.
Mark: Yes, GP attendances up as far as I know. Maybe they would have gone up even more without WiCs and MIUs? Not quite clear on your second point.
Joe: Yes, 'admissions' a typo - should be attendances. Will be corrected. Regarding where the 'real' problem in secondary care is, well yes, but we are writing about A&E here (which is still an important part of the acute system)!
Harry/Michael: On the 'supply induced demand' point, not all SID is bad of course. What we might have here is unmet demand at last being met through WiCs etc. Which is not necessarily a bad thing (though could argue if NHS treating things that people could deal with themselves is a good use of scarce resources, for eg). Equally, btw, SID by private sector also not necessarily a bad thing in health care or in other sectors; my demand for ipods was 'induced' by Apple's supply of the things - that's fine by me!

...Sorry, Sean, should have clarified that the 'subsequent attendance' data is rolled up into the 'total attendance' figures we use in the charts. The other bigger chunk of attendances are 'first' attendances.

Where is the problem? Public use of walk-in centres, urgent care centres, A&E etc reflects public approval of the availability of these services. No one goes to them without careful consideration first because the needs of the users are complex. The paternalistic condemnation of the public for using services 'inappropriately' is absurd and unethical. The services which the public vote for with their feet should be properly resourced and functional.

The argument that level 2 and 3 units account for the increase is surely overly simplistic? there was always small units and around this time remember widespread reconfiguation was occurring with many units merging leading to the downgrading of many.
The problem surrounding ED use has much more to do with the failure to manage unscheduled care in the community and GP referrals increasing, during a period of drastic cuts in bed numbers.
There is evidence that many of the ambulatory patients who should not need admission end up being admitted just because its not possible to get them home once they arrive in ED.
If we can't find ways to manage non emergency unscheduled/acute problems in the elderly population without ED referral for assessment as the default option, then further detioration is unavoidable

I note that there is a peak of about 5% higher than average each month probably correspending with the full moon, which indicates an astrological disturbance to our emotions when the sun and moon are in opposition

In reply to P Kerr. GPs manage most acutely I'll elderly in the community and only refer to ED when the patient is too sick to be managed in the community, or too confused to be safely left at home. the pressure in the system may be coming from an increasing number of sick elderly people living longer? maybe it would be good for ED consultant to spend a day with a GP and vice versa?

The GP academic Julia Hippisley-Cox has estimated that in 1995 the average GP practice of 8500 patients offered 20000 consultations per year, a figure that had increased to 34000 in 2008.Average GP consultation rates per GP have risen from an annual rate of 3.9 to nearly 6 in 2012.Perhaps more wooryingly, patients aged over 85 will consult on average over fourteen times each year.The causes are many but one could point to the greater emphasis on health promotion and secondary prevention , and this need has been exacerbated by the demands of the Quality and Outcome Framework.

Waht we are finding is that more of the work is now in the evenings and weekends, which is i feel partly to blame for the EM recruitment crisis, have you got stats showing in and out of hours workload please?

We knew this before - Matthew Cooke made this clear.
The data that is missing is the data about complexity.

A reasonable surrogate would be mean age of patients admitted, or mean age of patients arriving by ambulance.

The fact is that most patients who access Emergency Care do not have a meaningful diagnosis, so everyone is flying blind.

If we don't know what conditions bring patients to Urgent and Emergency Care, and what their final diagnosis is, it is impossible to make valid and reliable judgements as to the relative value of the different options.

Poor data = poor commissioning.

When looking at the A&E figures it's important to remember that the total number of A&E departments is diminishing, so although workload might appear fairly steady, the number of departments available to do it is falling. I'm sure they are feeling the pinch!
Chris Binn's comment #40434 is very apposite. Nationally GP workload is going up at a rate of approximately 10 million extra consultations per year, which dwarfs the A&E figures. There is no part of the NHS which is not struggling with changes in demography and demand.
To blame GP's for the problem might most charitably be described as over-simplistic.

The A+E stats are surely mirroring exactly what is going on in the whole of the NHS; an increased demand, fuelled by the ageing population with multiple health and social care problems (that are not met); increasing population numbers-especially of immigrants from the EU who cannot speak very good English and therefore take twice as long; and by the government who is constantly increasing people's expectations of the system.
I cannot work any harder as an inner city GP-often seeing 50-60patients per day as an individual doctor, and our practice team sees thousands of people per month. Most people, are seen and dealt with in primary care, but budgets are being cut by NHS England and targets for payments increased all the time.
Time to go home now; and i won't be working out of hours again on top of this huge in-hours commitment.

In many parts of the country GPs used to be paid for seeing patients with minor injuries attending their surgeries. These payments were stopped by the now defunct PCTs (not sure of the precise date, but about 2y ago?)and most GPs withdrew this side of their work, only treating emergencies on a Good Samaritan basis. I used to assess and manage a range of problems that my staff now deflect to MIU's & EDs BUT we don't keep any data on this, so its contribution to the ED burgeoning workload is unknown.
I can't see GPs taking back this work with the current trend of increased attendances for preventative and chronic care as fuelled (correctly most of the time) by QOF etc.

I love the way different people use and interpret statistics. These are gross numbers of attendances. Not rates. What has the biggest impact on the gross numbers? ONS says the population on 27 March 2011 was 56.1 millions for England and Wales. On the census in 2001 it was 52.4 millions. A rise of 3.7 million or 7%. The largest rise in any census period since 1801. Add onto people factors like increased morbidity due to the ageing population and that we are keeping people alive longer with greater burdens of disease necessitating greater levels of medical input, both acute and chronic. System factors like the new NHS 111 which promotes attendance at both gp surgeries and a&e departments ( I didn't realise we needed to stimulate new demand!). Our local OOH service works very well- we will wait and see it that is still the case after NHS 111 comes fully into place at the end of the month. Hospital factors, changes over recent years with hospitals cutting beds to safe money and pay for their PFI programmes means that gp referrals rarely go straight to the ward or surgical admissions units ( full) or medical admissions units ( full) so we are asked to send these patients to the a&e department. This clogs up the system. The poorly patients Stuck in a&e with no bed to move into for hours- a&e trolley blocked. Ambulances waits outside a&e hospitals is increasing i hear in the media- well that is no surprise to anyone surely. The bed status of my local hospital is red alert most of the time, they have over recent years created other alert statuses that are worse than red alert. The chief exec thinks he can trim upto 10 wards in the next 5 years or so. There are other factors too but i'll end there before it turns into too much of a rant and go find some GPs to blame even though it has nothing to do with them, it's part of their job description now you know.

I would be astounded if the national roll out of the 111 service does not result in a significant increase in attendance at type 2 & 3 units. In Oxford, where the service has been running for some time, an astonishing number of people with trivial problems are being directed to A&E by relativel untrained staff using cumbersome and unvalidated algorithms

This is an informative article, thanks for this. But what data are you using? There aren't any references unless I'm looking in the wrong place.

Patrick, thank you for your comment. The data sources have now been added to this blog.

Many thanks,

Jen

All very interesting. Everyone talks about 'demand' but not 'need'. These are two very different things, and it is disingenuous of the government to shy away from the need to educate the public about managing minor illness and injuries appropriately, while at the time time fuelling demand by hinting at the promise of unrestricted access and by bringing online more services (NHS Direct and 111) to heighten health anxiety among the population.

If we are in "time to think differently" mode, i was interested in Camilla Cavendish comments in the Times a couple of days back. She pointed out that the excellent University Hospitals Birmingham Trust had soaked up around 60% of the additional self-referral/999 activity across the City over the past couple of years, and argues that the marginal tariff for emergency admissions was perverse. In other words, if the public vote with their feet and pitch up at their local acute provider because they know it is any good and will look after them well, isn't it perverse to penalise the provider financially because the local population choose to use it. How about recognising that people have no confidence in alternatives and allowing the local acute care provider to grow its income by providing timely, responsive services that people want to use, instead of penalising it for the help seeking behaviour of the local public. Are we committed to genuinely patient centred services and choice? Or are we only committed to these ideals if they choose "care closer to home" or "ouside hospital". Stop demonising acute providers for failings in the rest of the system

David Oliver

Hi,

Really interesting stuff.

Not sure if appropriate here but can anyone point me to hypothesis based analysis on the drivers of increased AnE attendances based on a whole system modelling of unscheduled care?
I have seen GP referral management protocols, GP and ENP streaming, UCCS, tariff capping, QOF etc. I have tried historically to ensure data collection with common currencies across the whole local unscheduled care system through contracts but am met with the 'we do not collect that data", " we cannot afford to collect that data", contracts signed on an envelope without the detail and poor quality, unincentivised national data collections that aim to capture data that truly can't answer the questions we need to be posing. I understand providers will only collect mandatory fields in data sets that link to payments, when will we be more intelligent in our approach to this?
Any recommendations gratefully received as I see report after report and dashboard after dashboard and yet nothing really changes in content.
Any yes to publishing standardised rates data!

Reading this (interesting) post made me think about how I have changed how I use the NHS in the last 10 years. My local health centre typically has a 10 day wait for appointments, and refuse to do work like dressing changes. As a consequence, I no longer visit the GP (unless it's for the kids), but instead use a mix of a walk-in centre and a minor injuries unit, both of which are near the office. Both of them have opened in the last 10 years: as a consequence of poor local health provision, my "stats" have moved from GP to "A&E". Given that the walkin has a wait time of 2-3 hours (nearer 2 tbf), I also try and limit that interaction, relying in the chemist for minor problems. Overall I take the least problematic route!

With the release of figures that demonstrate a sustained rise in the elderly population, have you any further thoughts on this matter?

As I stated above in May last year

'The data that is missing is the data about complexity.

A reasonable surrogate would be mean age of patients admitted, or mean age of patients arriving by ambulance.'

I am trying to find A&E admission rates at ward level, specifically in three Boroughs of London: Redbridge, Camden and Brent and for the period 09/10 going forward (and ideally with London as the comparator). Does anyone know if this configuration of data exist?